Physiologically Informed Digital Auscultation for Pneumonia Detection in Long-term Care Residents
Abstract
Pneumonia is difficult to diagnose in older long-term care residents; multimorbidity and atypical presentations obscure signs, motivating operationally efficient objective testing. We analyzed multi-channel digital stethoscope recordings from 185 Japanese residents (73 pneumonia, 112 symptomatic without), using radiologist-confirmed chest X-rays and clinician diagnoses as supervisory signals that train convolutional neural networks, multimodal fusion, and channel-based variants with time-domain Grad-CAM interpretability. Models were evaluated with repeated patient-level cross-validation showing models with X-ray supervision outperformed clinician supervision (F1 0.729, accuracy 0.783 vs. F1 0.637, accuracy 0.711). Additionally, a three-channel selection protocol maintained performance (F1 0.736; accuracy 0.803), with two mid-thoracic sites ranking highest and Grad-CAM attention overlapping adventitious sounds. These findings indicate automated multi-channel lung-sound analysis can aid long-term care pneumonia diagnosis, with X-ray supervision being more reliable than clinical, and fewer channels preserving performance while lowering acquisition times.
Related
- Respiratory pathogens across rural populations: general practice patients with community-acquired pneumonia, residents, and goat farmers.
- A Case of Recurrent COVID-19 Pneumonia Following Rituximab Therapy: Treatment Informed by a Serial Transbronchial Lung Cryobiopsy, BALF SARS-CoV-2 PCR, and the Clinical Course
- ICU-acquired pneumonia beyond pathogen detection
- Identification of adventitious lung sounds in children with respiratory illnesses in Bangladesh using artificial intelligence and digital auscultation
- Preprocessing-Induced Ranking Reversal is Architecture-Dependent in Cross-Domain Pneumonia Detection
BibTeX
@misc{rasmussen2026physiologically,
title = {{Physiologically Informed Digital Auscultation for Pneumonia Detection in Long-term Care Residents}},
author = {Nicholas Rasmussen and Oleg Zaslavsky and Zih-Ling Wang and Hongyu Yu and Joelle Fathi and Kaibao Nie and Amil Khanzada and Tomoko Ito},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.27222},
eprint = {2609.27222},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.27222},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Physiologically Informed Digital Auscultation for Pneumonia Detection in Long-term Care Residents
AU - Nicholas Rasmussen
AU - Oleg Zaslavsky
AU - Zih-Ling Wang
AU - Hongyu Yu
AU - Joelle Fathi
AU - Kaibao Nie
AU - Amil Khanzada
AU - Tomoko Ito
PY - 2026
DA - 2026/09/23
PB - arXiv
DO - 10.48550/arxiv.2609.27222
UR - https://arxiv.org/abs/2609.27222
AB - Pneumonia is difficult to diagnose in older long-term care residents; multimorbidity and atypical presentations obscure signs, motivating operationally efficient objective testing. We analyzed multi-channel digital stethoscope recordings from 185 Japanese residents (73 pneumonia, 112 symptomatic without), using radiologist-confirmed chest X-rays and clinician diagnoses as supervisory signals that train convolutional neural networks, multimodal fusion, and channel-based variants with time-domain Grad-CAM interpretability. Models were evaluated with repeated patient-level cross-validation showing models with X-ray supervision outperformed clinician supervision (F1 0.729, accuracy 0.783 vs. F1 0.637, accuracy 0.711). Additionally, a three-channel selection protocol maintained performance (F1 0.736; accuracy 0.803), with two mid-thoracic sites ranking highest and Grad-CAM attention overlapping adventitious sounds. These findings indicate automated multi-channel lung-sound analysis can aid long-term care pneumonia diagnosis, with X-ray supervision being more reliable than clinical, and fewer channels preserving performance while lowering acquisition times.
ER -